Download Unexpected Suppression of Immunoassay Results by Cross

Editorial
Unexpected Suppression of Immunoassay Results by
Cross-Reactivity: Now a Demonstrated Cause for Concern
More than other analytical procedures, immunoassays
have afforded a wealth of knowledge about biochemical
physiology over the last 40 years. These assays have
posed major analytical challenges, which most likely stem
from the delicate balance between chemical equilibrium
and our ever-increasing quest for speed of analysis,
simplicity of procedure, and improved reliability. In attempts to enhance these criteria, we often face limitations
imposed by the fundamental principles of thermodynamics: balancing chemical equilibrium, the underlying kinetics, and that ever-present variable, time.
Cross-reactivity in competitive immunoassays can be
measured by various methods, including 50% displacement, equal displacement, or the gradient approach (1 ).
Typically, a cross-reactant is conceptually regarded as an
interferent causing positive bias in the assay results.
When an immunoassay is “rushed” (i.e., the signal is
measured well before reaching equilibrium), the effect of
the cross-reactant is enhanced, in essence increasing the
positive bias (2 ). That is the expected phenomenon. The
unexpected is the observation that cross-reactivity can
also lead to suppression in recovery (3 ), i.e., a negative
bias. In other words, the result obtained for a constant
amount of the analyte in a sample is lower when the
cross-reactant is present than when it is absent. This
indicates that rigorous characterization of cross-reactivity
requires studies in the presence of the analyte (1 ).
Suppression of assay results induced by cross-reactivity
was first described in Clinical Chemistry in 1996 for
digoxin immunoassays (4 ). In retrospect, however, previous observations hinted at this problem. Kanan et al. (5 )
reported that in plasma obtained from patients suspected
of having increased concentrations of digoxin-like immunoreactive factor, recovery of digoxin was lower in one
assay than in others. Additional immunoassays have also
been shown to be subject to suppression of results caused
by cross-reactivity (6, 7 ). In addition to digoxin-like immunoreactive factor, other cross-reactants, such as progesterone (7 ), digitoxin (8 ), or oleandrin (3 ), have been
reported to suppress recovery of digoxin in various
immunoassays.
A proposed mechanism for the observed suppression
of results lies in the physical design (i.e., architecture)
of the assay and is fundamentally based on the fact that
for binding of small ligands to antibodies, the rates of
association are comparable for primary ligand and crossreactant. It is the rate of dissociation that is different
and accounts for the lower binding affinity of the
cross-reactant molecule (2 ). The microparticle enzyme
immunoassay (MEIA) for digoxin (Abbott Diagnostics)
demonstrates a thermodynamic justification for this phenomenon. In this assay, digoxin in the sample (shown as
gray boxes in Fig. 1A) binds to anti-digoxin polyclonal
antibodies complexed to microparticles (MP in Fig. 1A).
An aliquot of this reaction mixture is transferred to the
matrix cell where the microparticle binds to glass fiber.
The matrix cell is then washed to remove the unbound
material. The enzyme (alkaline phosphatase) complexed
to digoxin serves as the tracer binding to unoccupied
antibody sites. The instrument measures the conversion of
substrate (4-methylumbelliferyl phosphate) to a fluorescent product by the enzyme. The amount of product
formed is inversely proportional to the concentration of
digoxin present in the sample. A cross-reactant (shown as
hatched boxes in Fig. 1B) such as progesterone competes
with digoxin for binding to the antibody sites with similar
association rate constants. However, during the wash
step, the dissociation rate for the cross-reactant is greater
Fig. 1. Mechanism of suppression of results by cross-reactivity.
(A), format when only the primary ligand (gray boxes) is present. MP denotes a microparticle with anti-digoxin antibodies attached to it. In this assay, the amount of
product generated as a result of enzymatic activity is inversely proportional to the concentration of the analyte in the sample. (B), when a cross-reactant (hatched boxes)
is also present along with the primary ligand, it will compete for binding to the antibodies. For a detailed description of the mechanism of suppression in this assay,
see text. Alk-Phos, alkaline phosphatase.
Clinical Chemistry 48, No. 3, 2002
405
406
Valdes and Jortani: Unexpected Suppression of Immunoassay Results
than that for the primary ligand and allows more unoccupied sites to become available to bind tracer. The
outcome of such a time-dependent competition is reduced
recovery attributable to increased tracer binding and
product formation. Therefore, the components necessary
to create the environment for this to occur are the presence of a cross-reacting substance in the sample and a
wash (or separation) step in the assay protocol. Short
assay incubation times can exaggerate the extent of interference. What is very relevant is that the phenomenon of
suppression caused by cross-reactivity is not limited to
any particular cross-reactant or to any particular immunoassay as long as the above-mentioned criteria are met.
To our knowledge, other explanations or plausible mechanisms have not been reported.
The clinical importance of this analytical problem was
not appreciated until recently. Steimer et al. reported in
Lancet (9 ) that misleading, subtarget concentrations of
digoxin measured by the MEIA assay (AxSYM) falsely
guided therapy and led to intoxication of a 71-year-old
patient. In this individual’s serum sample, potassium
canrenoate (given for ascites) and the metabolite of this
drug (i.e., canrenone) cross-reacted with the anti-digoxin
antibody, leading to reduced recovery of digoxin. The
extent of this suppression became evident when a sample
originally measuring 0.9 ␮g/L by the MEIA was reanalyzed by two other methods, both of which measured
digoxin as ⬎5.5 ␮g/L. They further confirmed similar
interference in serum collected from additional patients
on potassium canrenoate therapy.
In this issue of Clinical Chemistry, Steimer et al. (10 )
have further studied the clinical consequences of the
suppression of digoxin results caused by cross-reactivity
of several steroid-like compounds, including spironolactone, canrenone, and their metabolites. Although potassium canrenoate is not currently prescribed in the US,
spironolactone therapy has been increasing and is now
listed as one of the top 200 drugs prescribed in the year
2000 (11 ). Their investigation shows that the steroids
tested for cross-reactivity could increase, decrease, or
have no effect on the nine digoxin immunoassays tested.
Two of the three assays with decreased results use the
same reagents and assay design (i.e., MEIA) and have
been noted to have decreased results in the presence of
cross-reactants. Falsely low values were seen in ⬃4% of
patients and ⬎8% of intensive care unit patients. Alarmingly, three of four individuals with digoxin values in the
therapeutic range by the MEIA assay actually had toxic
concentrations of the drug in their serum. The study by
Steimer et al. (10 ) shows that negative bias introduced by
a cross-reactant is a serious clinical problem and that
several modern digoxin assays are prone to this problem.
It may be wise to reevaluate those immunoassays
currently on the market that by virtue of their design may
be susceptible to this kind of a problem. All of these recent
findings now point to potential concerns over clinically
important matters such as published reference intervals
for hormones (e.g., those with structurally similar endogenous compounds whose concentrations fluctuate with
time), dosing regimens based on therapeutic drug monitoring results, and variations in measured elimination
half-life of drugs.
In summary, the unexpected finding of a negative
interference from a cross-reactant was reported in 1996, its
mechanism was proposed in 1997 (3, 4 ) and subsequently
confirmed in 1998 (8 ), and the interference was shown to
cause a serious health risk for a patient in 1999 (9 ). Now,
in this issue, Steimer et al. (10 ) have taken this further and
demonstrate the broader impact of this type of interference. An important lesson is that unexpected phenomena,
when finally recognized as real, lead to new thinking, new
processes, and new approaches—all of which translate to
improved patient care.
References
1. Miller JJ. Valdes R Jr. Methods for calculating cross-reactivity in immunoassays. J Clin Immunoassay 1992;15:97–107.
2. Vining RF, Compton P, McGinley R. Steroid RIA— effect of shortened
incubation time on specificity. Clin Chem 1981;27:910 –3.
3. Jortani SA, Miller JJ, Helm RA, Johnson NA, Valdes R Jr. Suppression of
immunoassay results by cross-reactivity. J Clin Ligand Assay 1997;20:
177–9.
4. Jortani SA, Miller JJ, Helm RA, Valdes R Jr. Unexpected suppression of
digoxin values caused by DLIF. Clin Chem 1996;42(Suppl 6):S124.
5. Kanan R, Chan KM, Dietzler DN. Negative interference with Du Pont aca
method for measuring digoxin [Letter]. Clin Chem 1987;33:446.
6. Metheke ML, Valdes R Jr. Antibody specificity reduces interference by
endogenous digoxin-like immunoreactive factor. J Clin Immunoassay 1989;
12:115–21.
7. Jortani SA, Johnson NA, Brown GL, Valdes R Jr. Decreased recovery of
digoxin in serum collected from pregnant women. Clin Chem 1998;44(Suppl
6):A87.
8. Datta P, Dasgupta A. Bidirectional (positive/negative) interference in digoxin
immunoassay: importance of antibody specificity. Ther Drug Monit 1998;
20:352–7.
9. Steimer W, Muller C, Eber B, Emmanuilidis K. Intoxication due to negative
canrenone interference in digoxin drug monitoring. Lancet 1999;354:
1176 –7.
10. Steimer W, Müller C, Eber B. Digoxin assays: frequent, substantial, and
potentially dangerous interference by spironolactone, canrenone, and other
steroids. Clin Chem 2002;48:507–16.
11. Rx List. The top 200 prescriptions for 2000 by number of US prescriptions
dispensed. http://www.rxlist.com/top200.htm (Accessed December 2001).
Roland Valdes, Jr.1,2*
Saeed A. Jortani1
Departments of 1 Pathology and Laboratory Medicine
and 2 Biochemistry and Molecular Biology
University of Louisville School of Medicine
Louisville, KY 40292
*Address correspondence to this author at: Department of
Pathology and Laboratory Medicine, University of Louisville
School of Medicine, Louisville, KY 40292. Fax 502-852-7674;
e-mail [email protected].